Structure storing all metadata related to the resolution of a saddle-point linear system. A saddle-point system is depicted as. More...
#include <cs_param_saddle.h>
 Collaboration diagram for cs_param_saddle_t:
 Collaboration diagram for cs_param_saddle_t:| Data Fields | |
| int | verbosity | 
| char * | name | 
| const cs_param_sles_t * | block11_sles_param | 
| void * | context | 
| Main parameter settings | |
| Set of parameters to drive the resolution of a saddle-point system | |
| cs_param_solver_class_t | solver_class | 
| cs_param_saddle_solver_t | solver | 
| cs_param_saddle_precond_t | precond | 
| cs_param_convergence_t | cvg_param | 
| Schur complement approximation | |
| Set of parameters to drive the resolution of the pressure-related block. This is often a Schur complement approximation to B.A^-1.Bt | |
| cs_param_saddle_schur_approx_t | schur_approx | 
| cs_param_sles_t * | schur_sles_param | 
Structure storing all metadata related to the resolution of a saddle-point linear system. A saddle-point system is depicted as.
| A : Bt | where A is the (1,1) block |--—:--—| Bt is the (1,2) block which the transposed operator | B : 0 | w.r.t. B, the (2,1)-block
According t the choice of the solver, the saddle-point system is totally or partially assembled.
| block11_sles_param | 
Set of parameters used to solve (1,1)-block i.e. the A matrix. This is shared with the cs_equation_param_t structure.
| void* context | 
| cvg_param | 
Structure storing the parameters to know if an iterative process has to stop (convergence or divergence). These criteria are related the iterative algorithm used to solve the saddle-point system. This is the case for instance with an Uzawa or GKB algorithm.
| name | 
name of the saddle system or NULL (not mandatory). If NULL, the name of the (1,1)-block is used.
| precond | 
Type of preconditioner for the saddle-point system which is viewed as a 2x2 block matrix.
| schur_approx | 
Choice of the way of preconditioning the schur approximation
| schur_sles_param | 
Set of parameters used to solve the Schur complement if needed. This depends on the type of Schur approximation which has been chosen.
| solver | 
Type of solver to solve the saddle-point system If solver is set to CS_PARAM_SADDLE_N_SOLVERS, then there is no need to solve a saddle-point system.
| solver_class | 
class of SLES to consider (advanced usage)
| verbosity | 
verbosity (level of information displayed)